Question

I have a table, which links to another table specifying the the value of a key-value pair. The Value table then links to a table holding the name (key). The tables are defined (near enough) as follows:

---------------
Entity
---------------
EntityID
Name
Some other data

---------------
Value
----------------
ValueID
EntityID
Value

---------------
Key
---------------
KeyID
ValueID
Name

Each Entity can have many Values and each Value has one Key. Each Entity may not have the same number of Values. Basically, this structure is allowing me to add table rows dynamically, without modifying the Entity table.

I want to do a select * on the Entity table, using the Key.Name field as a column header, and displaying the appropriate Value.Values

Is there a dynamic structure available in SQL that I can do an insert into for each row in Entity, then return a select on?

I hope this makes sense!

Answer 1

Based on your description, it sounds like you want to PIVOT the data. There are two ways to do this with PIVOT, either a Static Pivot where you code the columns to transform or a Dynamic Pivot which determines the columns at execution.

Static Pivot:

select *
from
(
  select e.entityid, v.value,
    e.name EntityName, k.name KeyName
  from value v
  left join entity e
    on v.entityid = e.entityid
  left join [key] k
    on v.valueid = k.valueid
) x
pivot
(
  max(value)
  for KeyName in ([name 1], [name 2], [name 3], [name 4], [name 5])
) p;

See SQL Fiddle With Demo

A Dynamic Pivot determined the columns at run-time:

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX)

select @cols = STUFF((SELECT distinct ',' + QUOTENAME(name) 
                    from [key]
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query = 'SELECT entityid, EntityName, ' + @cols + ' from 
             (
                select e.entityid, v.value,
                  e.name EntityName, k.name KeyName
                from value v
                left join entity e
                  on v.entityid = e.entityid
                left join [key] k
                  on v.valueid = k.valueid
            ) x
            pivot 
            (
                max(value)
                for KeyName in (' + @cols + ')
            ) p '

execute(@query)

See SQL Fiddle With Demo

Both of these can be placed in a stored procedure. If you post some sample data for each of your tables and then the desired result, it would be easier to determine the proper query.